Differentiability of the value function and Euler equation in non-concave discrete time stochastic dynamic programming Articles uri icon

publication date

  • March 2019

start page

  • 79

end page

  • 88

volume

  • 8

International Standard Serial Number (ISSN)

  • 2196-1085

Electronic International Standard Serial Number (EISSN)

  • 2196-1093

abstract

  • We consider a stochastic, non-concave dynamic programming problem admitting interior solutions and prove, under mild conditions, that the expected value function is differentiable along optimal paths. This property allows us to obtain rigorously the Euler equation as a necessary condition of optimality for this class of problems.

subjects

  • Economics

keywords

  • dynamic programming; euler equation; envelope theorem